Methodology
Every ranking on this site comes out of a weighted scoring model that is fixed before anything is assessed. This page explains how that works. The criteria and weights differ per product class and are written out per class, listed at the bottom.
Where our figures come from
We work from primary sources: manufacturer documentation, official European product registers, and the specifications retailers publish. Every sub-score names the source it rests on, with a date. So you do not have to take our numbers on trust — you can look them up, and if a source says something different from us, we want to hear about it.
What we do not do is measure: we have no lab of our own. What we offer instead is that nothing goes in that we cannot evidence. Where a manufacturer does not publish a value, that cell stays empty and says «not stated» rather than carrying an estimate that looks precise. A figure without a source does not make it onto the page.
We earn a commission on referrals to retailers, through affiliate links that are marked as such above the first link on every page. Commission plays no part in the order of a list: the weights are fixed before any product is assessed, and the total is a sum anyone can recompute from the sub-scores and the weights.
What we do with owner reviews
For product classes where daily use counts heavily, what owners report in reviews at the retailers selling the product carries weight. What a manufacturer promises and what an appliance does after a few weeks are not always the same thing, and that gap is exactly what a buyer wants to know.
We do not collect those reviews ourselves. So we cannot verify that someone who wrote a review actually bought or used the product; the retailers decide whether and how they check that. That is why only a pattern recurring across several retailers counts here, never a single experience, and every score states which review pages we read and when. Any ranking that scores on reviews says so on the page as well.
How the order is determined
Each product is scored from 0 to 10 per criterion. The total is the sum of those scores multiplied by their weights. The order of the list follows from that total; it is not an editorial choice. If the figures change, the order changes with them. We cannot place a product higher than its score allows.
Every product class has its own model
Criteria that decide it for one kind of product say nothing about another. Noise matters enormously for something running in your bedroom overnight and barely at all for something you use for twenty minutes a day. So each product class carries its own criteria and its own weights.
Those models are numbered. Changing a weight or a criterion produces a new version number and leaves the old one intact; each ranking states which version it was built with. That keeps an earlier list checkable even after the methodology has moved on.